煤层底板突水危险性的PNN预测模型研究及应用

发布时间：2018-05-11 14:17

本文选题：底板突水 + 危险性　；参考：《中国安全科学学报》2015年08期

【摘要】：为准确有效地预测煤层底板突水的危险性,在分析大量观测实例数据的基础上,选取底板含水层水压、煤层采高、隔水层厚度、断层落差、煤层倾角和断层距工作面距离等6项指标作为影响煤层底板突水的初始特征指标。针对指标之间具有相关性的问题,利用主成分分析(PCA)法提取6项特征指标的主成分,将其作为概率神经网络(PNN)的输入向量,建立基于PCA的煤层底板突水危险性的PNN预测模型。选取21组煤矿实测数据作为学习样本,用于训练模型。采用回代估计法对模型回检。利用学习好的模型,预测另外4组矿井突水数据样本。结果表明,该方法有效降低了指标数据相关性,实现了降维,使PNN模型工作复杂度减弱。将该模型应用于工程实例中,所得预测结果准确率为100%。
[Abstract]:In order to accurately and effectively predict the risk of water inrush from coal seam floor, on the basis of analyzing a large number of observation data, the water pressure of floor aquifer, mining height of coal seam, thickness of water-separating layer and fault drop are selected. Six indexes, such as dip angle of coal seam and distance between fault and working face, are used as the initial characteristic indexes to influence water inrush of coal seam floor. In order to solve the problem of correlation between indexes, the principal components of 6 characteristic indexes were extracted by principal component analysis (PCA) method, and used as input vectors of probabilistic neural network (PNN), and the PNN prediction model of water inrush risk of coal seam floor based on PCA was established. 21 groups of coal mine measured data were selected as learning samples for training model. The model was checked by the method of back-generation estimation. Four other groups of mine water inrush data samples were predicted by using the model. The results show that the method can effectively reduce the correlation of the index data, reduce the dimension and reduce the complexity of the PNN model. The model is applied to an engineering example, and the prediction accuracy is 100%.
【作者单位】：辽宁工程技术大学系统工程研究所;
【基金】：国家自然科学基金资助(71371091) 辽宁省高等学校杰出青年学者成长计划(LJQ2012027)
【分类号】：TD745

【参考文献】